Lunar Dome Detection Method Based on Few-Shot Object Detection

Abstract Lunar domes have always been one of the important windows to understand lunar volcanic activities, but traditional geological dome identification methods are costly. This study attempts establish an automatic method for through FSOD (Few-Shot Object Detection). Since our previous research has trying automatically identify hills using recognition algorithms, team will try Few-Shot Detection this time. In study, researchers first obtained coordinates from list known and intercepted data need corresponding on DEM moon images. Subsequently, used existing train six algorithms compared their performance verify feasibility study. use AP50 AP75 evaluate model. Finally, found that can reach 0.76, 0.43. However, although does perform better than other methods, believe its accuracy is still far expectations. Admittedly, with YOLO v5, not significantly improved, verifies it feasible apply domes.